Cameras are often limited in their ability to produce sharp pictures by how steadily they can be held by the user. When a camera shutter remains open for an extended period of time, motion occurring during this open interval is visible in a snapshot. The visibility of the motion as a result of the combination of open shutter and motion is referred to as motion blur. Sometimes the introduction of motion blur into a captured image is purposeful, such as to capture the perceptual effect of high-speed motion more accurately or to provide a particular artistic effect. But for the photographer that desires a crisp picture, motion blur caused by “camera shake” presents an obstacle to that goal. Camera shake is primarily the result of rotational (e.g., pitch and yaw) motion of the camera.
Camera shake can be a limiting factor for exposure time, especially at longer focal lengths. Longer focal lengths magnify the image, and thus the perceived shake due to rotational motion is also magnified. A rule of thumb from 35 mm (millimeter) film photography is that, to avoid blur that results from camera shake, hand-held exposure times are selected to be less than the inverse of the focal length. For example, at a 60 mm focal length, the exposure should be 1/60 second or less. Considering the rule of thumb, there are various options to reduce motion blur. One option is to use a faster lens, which allows a shorter exposure time for the same scene brightness. Digital cameras typically use the fastest lens that is practical in terms of cost, size and image quality goals. Lens speeds of F/2 to F/2.8 (F referring to the F-stop, which is a calibrated measure of the ratio of a lens maximum aperture to its focal length, the inverse of which is an indication of lens speed) are typical. Faster lenses than this are often significantly more expensive and bulky.
Other approaches have been developed to address motion blur. One popular approach is active image stabilization of the lens system. “Image stabilization” refers to a process that attempts to stabilize an image on an image sensor or on a photographic film during the course of an exposure. In an image-stabilized lens system, a lens or prism disposed within the lens system is moved in such a way that the image path is deflected in the direction opposite the camera motion. The lens or prism is typically driven by two “voice coil” type actuators, which respond to signals generated by gyroscopes or accelerometers that sense rotational motion of the camera. As mentioned above, it is primarily the pitch and yaw motions of the camera that cause the image to move on the sensor. In general, the rotational motion is multiplied by the focal length of the lens, resulting in a shake that is “magnified.”
Liquid-filled prisms have been used for image stabilization. Such structures typically include two flat plates that form the front and back surfaces of the prism, surrounded by a flexible seal to hold the liquid in place. Actuators “squeeze” the prism by the edges of the plates, refracting the beam in the direction of the thicker side of the prism to counter image displacement on the sensor.
Moveable lens systems have also been used for image stabilization. In such systems, actuators shift the lens laterally, “decentering” the image provided on an image sensor horizontally and vertically. The beam is deflected proportionally to the power of the lens (positive or negative).
One problem with the image stabilization approaches described above concerns the limited space available within the lens system. For example, the moveable lens is typically located at or near the aperture stop of the lens system, which is a very “crowded” area in a camera, especially in compact zoom lens system designs. Additionally, the liquid prism approach is implemented using a separate, additional element to the standard lens system. Thus, the prism generally has to be fitted into the optical path. Further, lenses for these approaches are often specially designed to accommodate image stabilization, making them bulky, costly to fabricate, and complex in operation.
Another approach to image stabilization is leaving the lens intact and moving the image sensor. The image sensor may be fixed to a stage that is moveable in the x- and y-direction. The image sensor can be shifted by actuators in response to sensed motion, matching the movement of the image. One problem with this approach is that motion in the z-direction and in its tilt direction must be very carefully controlled, otherwise the image will not remain in focus. For example, out-of-plane motions of as little as 10 micrometers may cause some or all of the image to lose focus. An additional problem concerns movement of the sensor and the need for flexibly connecting the large number of signal lines from the camera control circuitry to the sensor.